Batboard

I'm posting this for a friend, but I will either answer or pass on all questions posed on this subject.

Current setup: stock MSF5000 VHF 120w station, digital-capable. 1 MHz split frequency. Sinclair 4-cavity BP/BR duplexer. Good jumper cables between the radio and the duplexer.

The duplexer has been tuned using just a Tektronix spectrum analyzer and tracking generator. There were no "pads" inserted between the test equipment and the duplexer. We can only assume that everything is 50 ohms. Passband loss is within specs. Reject loss is in the 80-85 dB range, which is not quite enough for a 1 MHz split, but with the serious filtering in the MSF5000's front end, it should be more than sufficient. To the best of my knowledge, the duplexers have always been tuned using the same equipment, and a dummy load has always been attached to the unused port. I don't believe that tuning with a long piece of coax has been tried.

Everything works just fine - no desense at all - if a dummy load is attached directly to the T connector (antenna port) on the duplexer. A signal can be injected into the receiver's front end (by inserting a wire into one of the tuning holes) and the receiver will open squelch at a certain signal strength regardless of whether the transmitter is on or off. During this condition, the signal level on the metering pin 2 remains constant around 8-10 regardless of the transmitter power output, all the way up to around 175 watts (yes, I know, way over the limit but it's never kept there for more than a few seconds).

Next step is to insert 100 ft of new 1/2 inch Heliax between the duplexer and the dummy load. Now it takes between 12 and 20 dB more signal to keep the receiver unsquelched when the transmitter is on, than when it's off. We're considering this to be 12-20 dB of desense. When the transmitter is ON, with no signal feeding the receiver, the signal level on the metering pin 2 goes up over 30 and varies if the transmitter power is adjusted up or down. The amount of receiver signal that's needed to overcome the transmitter noise is directly related to the transmitter power - run it down at 10 watts and only 2-10 dB more signal is needed to open squelch with the transmitter on.

Now the fun part. So far he has had this ongoing problem with two other duplexers, one of which was 6 cavity, several MaxTrac radios, some GM-300 radios, and even different coax on the antenna port and between the radios and duplexer. As soon as a length of coax is added to the antenna output of the duplexer, the desense goes through the roof. We all figured that an MSF5000 would cure his troubles... NOT.

He has put a spectrum analyzer on the receiver input and observed what's coming down the pipe. The transmitted signal shows up around -40 dBm, which seems to make sense, as a 100w signal would equate to +50 dBm and the duplexer should attenuate that somewhere between 80 and 90 dB. He has scanned around looking for any other energy in the region of the receiver's frequency, plus or minus 21.4 MHz (twice the IF) and found nothing. He can see nothing at the receive frequency down as low as about -120 dBm (that's the noise floor on his spectrum analyzer).

It almost appears that the transmitter, or exciter if you can call it that, in the RF tray, is getting into the receiver from some path other than the front end. But only when some length of coax is added to the output of the duplexer.

It doesn't matter if he connects a dummy load or an antenna to the end of his various lengths of coax on the antenna port - he still gets about the same amount of desense.

Between my friend, Nand, and myself, we're out of ideas. Time for some new brains to attack this problem as it has driven us nuts for several months.

Thanks to all.
Bob M.

Wow! Isn't radio fun? Have you guys pulled out all your hair yet?

Different radios, different duplexers, different coax, new heliax [100% shield] SAME problem?

What's left as a common denominator? The Tektronix?

You didn't say that any other instrument was ever tried for the duplexer alignment. You could start with using pads during the alignment and then if that doesn't work - try another instrument.

IF that doesn't work - you may have to sacrifice a couple virgins to the repeater gods - must have done something to offend them.

One other thought - buy or construct a proper ISO-T for conducting the desense test - like the Bird 4275-025 or a Microlab/FXR HX-A17

This sounds like some of the stuff that has happened to me. Have you used a different cable than the hardline? Try a hundred feet of lmr400. It seems to me that since everything is OK out of the duplexer, you can eliminate arching within the duplexer. Have you checked the return loss of you hardline/dummy load combination? I am leaning toward either a bad connector on the hardline, or bad hardline. You should be able to see something at the receiver port with a spectrum analyzer though regardless of what is causing the noise. I had a problem with a MASTRIII that was beyond weird. But, we are looking at your problem now. Let us know what turns up.

Bob said: "...and even different coax on the antenna port and between the radios and duplexer..."

That would appear to rule out the Heliax.

In any case Heliax would be 6000% preferable to LMR-400 or any other braid-over-foil cable. If he had said he was using LMR-400, getting rid of that would be the very first recommendation.

This subject has been covered numerous times on the various repeater lists. Here is a quote from Kevin Custer:

"...It is the consensus of the masses that ANY shield over foil
type cable not be used for duplex service. Andrew Heliax, Superflex,
RG-400, RG-142, and RG-214 are the best. No problems with those cables
in duplex service.

Kevin Custer"


That does bring up the subject of the jumpers between the duplexer and the transmitter and receiver - these must be double shielded or superflex type cable.

In suggesting that they try LMR400, I only meant it in the sense that they should try something different to rule out the hardline and connectors. Obviously, hardline is better than LMR400. If someone else doesn't like it, that's OK. I'll continue to use it until I have seen a reason not to.That said, I know of several repeaters using LMR400 without problems. Something is causing the desense here. I have kicked myself several times when I assumed something to be good. This is a very interesting problem. I'll bet that the answer will be very simple once it is found.

OK, first round of replies. Thanks for the suggestions.

The original coax going to the antenna out on the roof IS LMR-400. A dummy load was tried at the end of this and we still had desense. Another 100 ft hunk of LMR-400 was unrolled and exhibited the same problem. A brand new 100 ft piece of Heliax was obtained and exhibited the same problem. Any coax attached to the output of the duplexer seems to be producing desense.

Pads were not tried because I have some and my friend does not.

I also have a return loss bridge. My friend does not.

I have an HP and an Agilent spectrum analyzer. My friend only has the Tektronix unit.

We've both successfully tuned UHF duplexers, one cavity at a time, with less sophisticated equipment (i.e. a receiver and a signal generator). The VHF duplexer seems to be tuning up correctly. Slight tweaking of the notches, passbands, and coupling, done while connected and in service, seems to have no effect on the level of desense.

At one time he had 40 inch well-shielded cables running between the radio and the duplexer. This ends up being about 1/2 wavelength (mechanical) which we've learned is a "Martha Stewart" bad thing, so other cables were cut which were closer to 60 inches. At first it seemed to improve things, but at the same time the receiver lost sensitivity due to new whiskers growing inside the tinned cavities. Once the whiskers were cleaned out and the receiver was able to detect a signal down around 0.1 uV, the desense got even worse (went from 12 dB to 20 dB).

We HAVE pulled our collective hair out. The items that have not been changed are the test equipment, the person doing the aligning, and until recently, the interconnecting cables. We even looked for a strong AM station on 1000 kHZ, but the closest one is several kHZ and more than 10 miles away, even though it is running 50 kW.

I hope I answered all the questions in this first round. What else is left to try? We're trying to figure out if it'll be cheaper to fly me and my test equipment down to the owner's house and back, or ship the repeater and duplexer up to me and back.

Oh, there is a "peripheral box", a.k.a. filter with T/R relay, in the MSF5000. We've tried running the output of the power amp through that as well as around it, and it seems to make no difference on our desense problem.

As Dr. Mom would say on TV, "Next?"

Bob M.

Once you have the duplexer on the radio, how are you driving your stimulus signal down into the radio? Are you transmitting over the air, and picking the signal up, or are you driving the signal into the system?

What you need is a service monitor which can both terminate the 100W transmit signal as well as generate a stimulus signal for the radio. That way, you can really measure the desense "properly" - through the duplexer.

I would be suspicious of injecting a signal into the radio in the fashion you suggest - you may be overloading one of the first RF gain stages with the TX signal, and your method of stimulating the radio may be bypassing that overloaded first RF stage.

Otherwise, could you have a mismatch between the antenna and the output of the cans? If the shield of the feedline were radiating due to impedance mismatch at the antenna, that may be enough to leak back into the radio.

Does the duplexer response change if you change the load on the duplexer antenna port from a dummy load to a dummy load at the end of the long run of Heliax? You should not see any difference between the TX to RX port when viewed like this in the tracking generator setup.

Nand.

There have been a number of threads on here about duplexer tuning and desense problems. You may want to search for some of it.

First and of the most important item to always remember, you need to keep the impeadance of the system constant. If you don't use series padds, you can't do this. When you change the cables, your adjustments will go down the drain. A 6 to 10 Db pad on all 3 points of the duplexer is needed. Put the pads in series with the duplexer and your test equipment.

Inject from the common T point and measure at the RX or TX ports. You will need a strong signal to get through all 3 cavities on each leg of the duplexer. You may even want to use a receiver when your tuning for the notch. If your signal generator doesn't have enough guts to generate the signal, you could use a portable on low power.

The way you tune will depend on the type of duplexer you have. Problem is you can't do one cavity at a time and then just hook the cables back up. It won't work as your finding out.

You may also want to make a capacitive coupling device to inject the receive signal into the system once the cavities are tuned. The simple way it to take a t coax connector and remove the side center connection. Then connect the signal generator into this modified port. The rest you already know how to do. Go look in the Radio Amateur Handbook on how to make this modified connector if your not sure from my instructions.

Jim

2nd round of answers.

As far as I know, the sig gen was originally fed into a piece of wire, or something equally bad at radiating, when the duplexer was feeding coax that had an antenna on the end. This got a signal into the receiver via the antenna and through the entire system. I don't know if that's been done lately.

When the dummy load is on, an antenna can't be used on the sig gen, so a short cable, similar to the MSF5000 tuning cable, is being used to inject the sig gen's signal into one of the tuning ports of the front end. The amount of signal needed to break squelch with the transmitter off vs with the transmitter on, is giving us the 12-20 dB of desense measurement.

I'm pretty sure that the current 4-cavity duplexer is being tuned with all the cavities connected, as the notch is well within the capacity of the sig gen and tracking generator. On my Cellwave 6-cavity UHF duplexers, each cavity yields 40 dB of notch, and you really can't see the 100-120 dB of rejection well enough to tune all the cavities while connected. But that's not the problem here. We've both read almost all the duplexer tuning info available - I think I even added some to what's here on the BBS.

Also, ALL 5 of the duplexer's T connectors have been replaced. One was found to have a faulty center pin connection and was causing noise and arcing when it was banged.

I have 6-12 dB pads. I have a Coaxial Dynamics 10-80 dB signal tapoff which will also inject a signal. And I have an Eagle-1st return loss bridge. The person with the problems has none of this, and so far has not requested that I ship all this stuff to him. I agree that it would make the desense measurement a bit easier since we could inject directly into the antenna connection of the duplexer.

I don't believe any difference was noted in the response curves with different connection configurations, but I'll attempt to verify that. I also don't remember if the system was tuned with the antenna and coax attached - if it was, it made no difference otherwise the problem would have been minimized or even eradicated.

Point well taken about using the service monitor, but we don't have one available, so we're doing whatever is possible without it. I suppose some of the transmitter's signal could be getting into the receiver when we try to inject a signal, but we get the same desense condition (i.e. a weak signal will cause the repeater to key up, drop out, key up, drop out, etc) with the coax and antenna attached, so the desense is real and not something that's being manufactured as a function of the test setup.

I'm going to try and speak to the equipment owner tonight and make sure he's following up on this thread and can provide some answers and/or make corrections to anything I've said here that's inaccurate.

Ghosts and Goblins approaching ! Gotta sign off now !
Bob M.

gotta go with wowbagger on this one...

you're testing a duplexer, but bypassing the duplex function of the device
technically, you are using two antennas one TX (the length of cable) and one RX (the wire used to inject signal)

...Z

Well, perhaps, however...

Technically a piece of Andrews 1/2 inch superflex Heliax should NOT radiate the transmitter signal, so the fact that we don't get any measurable desense with our test setup (i.e. injecting a signal into the receiver's front end via a piece of coax) when a dummy load is attached to the antenna port of the duplexer, compared to the 12-20 dB of desense we see if we put that 100 ft piece of Heliax between the duplexer's antenna port and the dummy load, are we injecting any more transmit signal into the receiver this way than without the 100 ft piece of coax? And it doesn't seem to matter if we replace that coax with another 100 ft piece of LMR-400 with an antenna connected to the far end. Obviously the transmitter's signal, or a spur/harmonic/undesirable emission is clobbering the receiver's ability to properly hear the weak signal we're injecting into it, or another distant station is providing.

If everything is a true 50 ohms, and I agree that it isn't in the real world, then throwing a piece of real good coax between the antenna port and dummy load should have no effect, unless we're hitting some voltage or current peak due to the line length being just the wrong multiple of a wavelength or portion thereof.

It seems logical to me that if I called Sinclair and said "Build me a 4-cavity duplexer on these frequencies", they'd do so and tune it up using factory equipment and 50 ohm loads. I should then expect to be able to just connect it to my MSF5000 and antenna/coax and NOT have to retune or just touch up the adjustments, unless something in my equipment is extremely far away from having a 50 ohm impedance. Yes, I might be able to tweak slightly better performance out of it, since the connecting coax would be different from what Sinclair used, but it should at least be in the ballpark and not give me such a hard time. Maybe I'm expecting too much, but I don't think that the Sinclair instruction manual mentions retuning or touch-up when the unit is put into service. If that were necessary, they wouldn't even have to bother tuning it up at the factory - just let the end user do it to match his/her situation.

So many items have been replaced already. What's left? The sig gen, spectrum analyzer, tracking generator, the ham trying to make it work, and the geographical location?

Bob M.

Whether the signal is injected through one of the front-end cavity injection ports of the MSF5000 or at the antenna port using an isolated T connector should not make a noticable difference. Either way you likely have around 30 dB of isolation between the signal generator and the signal path of the receiver. It certainly should not upset the rejection ability of the duplexer.

But I believe the problem also shows itself when a strong enough signal is radiated near the MSF5000 without actually making a physical connection between the MSF5000 and the signal source. There apparently is a 12 to 20 dB difference in signal needed depending on whether the TX is on or off.

Didn’t your friend mention that there was a difference in desense when he used another length of cable between the MSF5000 TX port and the duplexer? This would indicate an impedance miss match some place.

Nand.

I don't recall if he tried a shorter piece of coax, besides maybe something 3 feet long. It is something to try. Since the desense is directly related to the transmitter output power setting, perhaps it can also be reduced by shortening the coax length.

Let's just say for argument that there are three possible outcomes of trying another, shorter piece of coax. The desense could stay the same, it could get better, or it could get worse (and we don't even want to think about how much worse it could get). What does this suggest? It may be that the shorter piece of coax is closer to a critical length (i.e. multiples of 1/2 wavelength) than the 100 ft piece.

I am hoping that the owner of all this equipment will read all these messages, try everything that has been suggested if not already done, and report back the findings so we can rule out more possibilities.

Bob M.

The point was that if the cable length between the transmitter output and the duplexer makes a difference in desense, then the duplexer input impedance is not the same as the cable impedance. In this case that would be 50 ohms.
Or if the input of the duplexer TX port is 50 ohms, it may not be 50 ohms for spurious and other noise coming out of the transmitter. If that is the case, the length will affect the amount of noise passed on.
When there is a proper match and no noise, than the cable length at that point should not matter. And if there were spurious noise coming out of the transmitter, then the length that would be optimum would be dependent on the frequency of this noise component and the impedance that the duplexer presents at that frequency. The optimum length would not have any relationship with the operating frequency.

Nand.

I agree with most of that.

However, why do duplexer manufacturers suggest/specify that the length of the coax from the duplexer to the radio should be an odd multiple of 1/4 wavelength, i.e. 1/4, 3/4, 5/4, etc.? Sure sounds like the length is somewhat critical, or at least important, to the overall operation of the system. And why is an even multiple of 1/4 wavelength, i.e. 1/2, 1, 1-1/2, etc., such a bad thing? Earlier in the month I was at a VE test session and posed the cable length question to the hams present. The immediate and unanimous answer, from four of them, was that the cables need to be an odd multiple of 1/4 wavelength. And some of these people also have GROL certificates (the replacement for the old FCC First and Second Class Radiotelephone Licenses). There must be some truth to this.

I believe that when the interconnect cables were changed from 40 inches to something closer to the odd multiple length, the desense initially appeared to be better, but on closer observation, whiskers had regrown in the front end and desensed the receiver by itself. Thus the entire system was very insensitive and desense measurements appeared to be better. Once the receiver was repaired and sensitivity was back to normal, the same 12-20 dB of desense was observed.

In a perfect world, everything would be 50 ohms. I expect that the equipment we're dealing with here is close to that ideal value, but certainly could be off some amount.

Latest info is that the reflected power, measured anywhere that it's possible to do so, is zero or barely visible on a Bird wattmeter. Now this doesn't necessarily mean that everything is 50 ohms, but there must be enough impedance matching going on to allow 100-150 watts of RF to leave the transmitter yet see nothing or a fraction of a watt get reflected back from any load. I will admit that introducing the wattmeter into a length of cable will change that length, and (again) ideally one should also have an additional piece of coax attached to the meter so the total introduced length is a multiple of 1/2 wavelength. But unless the line length is extremely close to a 1/2 wavelength multiple right now, such that the introduction of 4 more inches will matter, then the cables need to be adjusted or replaced.

The owner of the equipment is going to go through this entire post and try any and all suggestions possible before replying with his findings. Another BBS user suggested adding a circulator to the transmitter output, and he offered to send one to the owner of the equipment. I thought that the MSF5000 already had a circulator in the power amp, but perhaps that's only in the UHF models.

I also asked the owner to do a spectrum analyzer sweep of the peripheral box, which we believe contains a low pass filter plus the T/R relay. Then we will see what the components are attempting to filter.

Bob M.

Ok, Bob has finally dragged me out of the shack to post some additional details here.

First, thank you Bob for all the help so far with this pain in the neck problem and all the time and effort you have put in to help me.

Thanks also to the folks here who have already lent some spare brain cells to the issue.

On to the problem, with a little more background.

The most unusual, and perhaps the key to this problem is that over the last 8 months I have changed everything that should have been a factor as follows.

3 different duplexers, one of which was a 6 cavity PD.
Many different radios from GM-300s to ham rigs now to the MSF-5000.
3 different coax lines lastly the Heliax.
3 different antennas.
All different interconnecting cables of varying lengths and types, too many to even recall.

So, my approach has been to upgrade everything to the best available thinking I must have a weak link causing the de-sense, so far with no luck.

The test that I ran several weeks ago which really has me stumped is where the dummy load is on the output of the duplexer with whatever power output I want and the RX is as clean as can be, not any hint of de-sense, nothing at all. The "problem" can be seen clearly on meter position 2 of the metering panel. The normal noise floor is one reading, when the transmitter is keyed with the dummy load at the output of the duplexer the meter does not move at all. When a transmission line with dummy load or antenna is connected that meter reading increases indicating de-sense. If the squelch is open you can hear "typical" de-sense.

As for questioning the transmitter, along with the fact that everything has been replaced I stuck in a GM-300 for a transmitter just to once again verify that it does not seem to be related to the electronics. Same performance with the GM-300. De-sense with transmission line and clean as can be with the load at the output of the duplexer.

Throughout all this I have made many measurements with my Bird at various places in the system. Output of the repeater, output of duplexer, at the antenna or load at the end of the coax. Always zero reflected power, again the meter does not even move on reflected power, looks like a very good match throughout the system. The dummy load is a DB products 100w load.

I myself started to question my tuning and test equipment, it's only natural to do so since I have replaced EVERYTHING else. Tuning a duplexer is just not that hard. I have been doing so for decades. In fact over the last several months I have tuned 4 other UHF units for other Ham friends who have them in service with no problems, this sort of lends some credibility to my equipment and processes here. However, it has to be something going wrong.

If it were not for the fact that it is as clean as a whistle with the load at the duplexer output I would just say the the 85 or so db isolation I am getting is just not enough, however with the 6 can PD unit it was the same so I really can't conclude that.

As for use of pads in tuning, I am myself skeptical of the value. I don't know what the theory is behind this, if they would be so critical in tuning how would the system work without them once they are removed? It's just not that hard to align cavities. He said... :)

Again I am stuck in the mud here, it must be something, and as Bob mentioned the only thing that has not been swapped out is me, however I have tuned other units recently with perfect success.

I don't think any different method of accurately measuring the de-sense would add useful data, open to suggestions here, and everywhere for that matter. The de-sense is there in a closed system with transmission line and gone when the transmission line is removed. I assume this removes my local RF environment all together, which can be a big factor.

Oh, did I mention I live across the street from a particle accelerator, kidding...

Znarx comment on test setup, yes I see your point however that reflects only one test configuration and I don't think contributes to the overall problem.

Nand, many different cables have been tried between repeater and duplexer. At one point I thought a different length did make a difference, it did not. The problem was the T connectors which have been replaced with proper ones, de-sense is the same as with old ones. With the bad T's it was just flaky operation in general. Duplexer re-tuned after connector replacement naturally.

Thanks again for everyones time on this issue. Clearly there is something very screwy here so something MUST be the cause, I just don't know what.



John

Let me ask a question that hasn't been asked or covered yet. What is the location and tower like? Is this a clean no other radio location or is this a common comm site with all sorts of other radios? Is the tower new or old?

Let me start with the tower. If this is an old tower, there probably are all sorts of loose and rusting hardware. These bad metal to metal connections are diode joints and cause all sorts of intermod and desense problems.

What is the ground like at the site? Your coax cable shield needs to be tied to the site ground as well as the radio cabinet.

Are there any other transmitters at the site that are on the air most of the time? Like FM stations, TV stations, paging transmitters. They all cause undue hair pulling to solve intermod / desense problems.

I have seen the local oscillators in other radios nearby to cause a desense problem. These are problems that are hard to locate.

Even found that one of the TX multiplier stages in a UHF transmitter to cause an intermod problem with a 2 meter ham repeater. It took several days with a calculator to loacte this problem. Solution was to obtain another RX crystal for the ham repeater and retune the RX for the other side injection.

By any chance have you looked at the cables from the duplexer to the radio? These need to be of the best quality shielding that can be obtained. As well as the rest of the duplexer cables. You can't use single shielded cables here. Double shielded or heliax type cables only. Can't use RG-58 or RG-8 type cable here.

Last but not least, by any chance is the duplexer a DB Products unit? If so, is it more than a couple of years old?

Jim

John,
From all our previous conversations with you and Bob, it looks like you looked and tried pretty well everything that can be done but the results are always the same. Just like a bad dream. Here are a couple of things I would like you to try and some observations as well.

There are circulators in the PA with an associated load, I believe. If anything went wrong there, then the PA would not see the proper termination and could cause all kinds of noise. Adding an external circulator will not fix this problem without removing the faulty one preceding it.
I really would like you to get a single pass cavity and see what difference it makes if placed in the TX output. A cavity there goes along way to ensuring that the TX signal is clean.

If there is no desense at all when the dummy load is place directly at the duplexer output but there is desense if there is a Heliax run between the dummy load and the duplexer output, then the impedance of the dummy load / Heliax combination is not 50 ohms for some reason. Even though you measure no reflected power, there may be lots of reflected power at a much lower level at spurious and noise frequencies compared to the 100 watts of the desired signal. Adding short length of cable to the Heliax may show a difference in desense in that case but not cure the source of the problem.

I would also like to know the results of a measurement with the tracking generator setup across the duplexer TX and RX ports and the dummy load on the duplexer antenna port. Does adding the Heliax in line with the dummy load now make a difference in the measured isolation? It should not!

Nand.

Hi:

For what it is worth, and I am no expert here.

As a field tech , I get dispatched to mountain tops
all the time, sometimes to chase down problems of
this type.

A couple of issues to start with:

1. You really need to take a look at the Transmitter
with a good RF Spectrum Analyzer !

Analyzers in Service Monitors although they are
handy, they are no match to a Tektronix 492
which has superior resolution

After you know your transmitter is clean, and
hopefully it is, take a look at nearby transmitting
antennas, do you have a Paging Transmitter
nearby, TV or FM Transmitter nearby ?

High Power Transmitters can be a real hassle to
deal with.

Always remember, you have a Antenna out
there, it loves to pick up RF ( from anywhere )
and sometimes " RF Noise " can get into "your " Transmitter
( Solid-State Transmitters ) are real victims of this, and
they can ( under the right conditions ) Mix and amplify
the Noise Floor.


Also, do NOT OVERLOOK the torque Settings of Hardware
on Shields, and covers in the MSF....Yes they have special
Torque Settings, and they should be observed !

RF is like water, it leaks everywhere.

Next level:

What is the Noise Floor Level Look Like ?

Assuming its typical, and not unusual, then I will try
a " Small " Alternate antenna, even a good Mobile
Antenna ( with Low VSWR ) can be used.

See if the Desense is still there.

The Common issue as I understand it, is that you
have very little ( if No ) desense when a Dummy
Load is used, but when connected to the Antenna
System, it starts to have a problem.

When running " High Power" one can have all types
of issues, and you also can reduce the RF Power Level
and see if the RF Output has a impact on the RX Desense.

If you reduce the RF power by 3DB, then your Desense
level should drop accordingly.

Lets say it does not drop......Well, you have a internal
issue with the MSF somewhere.

Some Tech Notes:

I have seen Bad Transmission Line all the time, and a
TDR or suitable cable analyzer is required to locate
where a suspect problem might be.


May " Look Like " low VSWR, however, a Analog Meter
such as a wattmeter is only a basic tool.

What we do, is measure the Power Going into the cable
@ the Source Location, then, measure the Power going
into the Antenna @ the Load Location.

Taking into account the Length of cable, one can
make a general assessment if the Cable and antenna
may be OK.

OLD used Antennas can be a problem as well ! and
High Power Transmitters " Hate " Noisy antennas !

They can even heat up the Cable, and even melt
the interior parts of a antenna.

Noisy antennas are defined as antennas that may have
cold solder joints inside them, and when they move
around, ( as with wind ) they can cause all kinds of
problems , not to mention when RF tries to Jump across
the Cold Solder Joints.

Anyhow, just food for thought, next trip to the Hill, take
along a New Antenna, mount it outside the Building some
where, run a piece of RG-214, or any other suitable cable that
is well shielded, and see if you have the same problem.

Monty

I should point out here that John’s problems shows up without any antennas at this point. He is using a 100 watt dummy load for all his tests and still has the problem. It wouldn’t hurt to try a different dummy load though.

Nand.

Both Monty and Nand brought up the next point that I was waiting for more info before I said anything. The spectral purity of the new all solid state transmitters is junk compared to the good old tube days.

Most of the newer all solid state PA systems are wide band. There is none or very little tunning provided in the transmitter outputs today. As a result, the requirement of the PA to be broad band causes the noise floor to come up on these transmitters. Some are more noisy than others. The PA just amplify's any and all junk that gets inputted to it. It's just a broad band noise amp.

I will mention it here as no one esle has, a poorly tuned stage before the PA will cause some of these same noise issues. You may even have to take a Spectrum analyzer to look at these stages before the PA.

One simple way to try to help clean this up is to put a pass cavity on the output of the transmitter. When using a duplexer, it has to be placed on the TX port. It can't be placed on the receiver or antenna port.

You could also place a notch cavity between the duplexer and the receiver. Tune the notch for the TX frequency. This will add to the isolation that the duplexer should be providing. Just remember that the cable lengths are important. It will also add insertion loss.

Like Nand pointed out, you may have an isolator issue with the MSF5000. Remove the internal one and obtain another and place it on the TX output. Don't forget to have a load on the termination port that has a power rating of what the TX is putting out. This isolator will prevent anything coming back down the cables into the TX from mixing in the PA.

It may sound like you have been hit with many issues to go look at. Problem is, any one of them or a combination of any could cause just what your seeing. It just takes trying every one of them till you find the problem. When you find it, your probably going to kick yourself for going around it for so long. It will more than likely be some simple thing that you just over looked. Like that couldn't be causing the problem.

Something as simple as corrosion on the antenna connector where your trying the dummy load could be the issue. You may want to take apart both ends of your feedline and look real close to the inside of the connectors. If when you take off the bottom connector, water comes out of the 7/8 cable or connector, you could have a bullet hole in the cable. I think you said your using 1/2 inch. Only a real good TDR will show that up.

If you can get your hands on a Site Master, use it to look at your feedline. Blow the scale way up so you can see every bend in the cable. It should also show you if there are any cable clamps that got over tightened all the way to the antenna.

Try taking the antenna off and putting on a 10 DB pad or a good 50 ohm termination. You should see a fairly flat sweep on the Site Master. Maybe a small bump for the connector. Anything else or any large spikes or what ever, need to be looked at real close.

Keep us all posted as to the steps you take and what you find.

Jim

All the points that have mentioned antenna systems are nice to know, but they have absolutely no relevance to the current problem, as John, Nand, and I have stated numerous times.

When the output of the duplexer is fed directly into a dummy load (NO coax), the receiver has NO desense problem.

When the output of the duplexer is fed through 100 ft of Andrews SuperFlex 1/2 inch Heliax into a dummy load, the receiver has 12-20 dB of desense.

Raising and lowering the transmitter power directly raises and lowers the amount of desense, or the signal level needed to open the receiver's squelch, as compared to the level when the transmitter is off.

Antennas inside or outside have absolutely no effect on this problem since they are about 60 feet away and not connected to the system. Unless John got that special radiating form of Heliax, there should be no radiation coming from a tight, new piece of coax. Ditto with the LMR-400 that he also has.

He does have a fairly good Tek spectrum analyzer but I don't know if he has found a way to look directly at the transmitter's output signal without blowing up the SA's front end.

I have been arguing for months that these new transmitters are wide-band and untunable, unlike the 1970s vintage ProgLines and older units. The MSF5000 PA is wide band. The IPA is wide band. There are no adjustments at all for the transmitting signal in the RF tray.

It might also be interesting to introduce a cavity (bandpass or reject) between the IPA and PA to see if that helps anything.

Another test that might prove interesting, even though time consuming, would be to reconfigure the duplexer for three cavities on transmit and one on receive, so the receive signal gets notched even more than it does right now. I just don't know if the receiver's front end can stand a +10 dBm signal on the transmitting frequency. That's pretty potent and might cause desense all by itself.

But so far he can't see an offending signal at the receiver's antenna connection with the test equipment he currently has available.

Bob M.

Lets try some different trains of thought...

I assume there's an actual range problem with the repeater using the antenna with 100ft of Heliax when he goes driving around with a couple of mobile radios for testing? If so, that would indicate his desense testing method is likely sound, instead of suspect.
Has he tried reprogramming the repeater with a 5MHz offset, and retuning the duplexer to the same, just for testing purposes? That might at least tell us the duplexer is doing the job at the 'usual' frequency spread. I have found out the hard way tuning a duplexer for VHF frequencies only 1MHz apart can be tricky without the proper equipment. The measurements look ok, but it doesn't do the job. By the way, Sinclair's standard 4-cavity duplexers only offer 70dB of isolation. The standard 6-cavity offers 80-85dB, which is still not quite good enough, as has been pointed out. Most times when I've seen VHF repeaters with that close of an offset, they've used a series of multicouplers, as opposed to a duplexer.
If he eliminates the duplexer, hooks his test equipment into the receiver, and puts the 100ft antenna run directly on the transmitter, does he still get desense? If the antenna is up a good 50-60 feet and outside, that should provide plenty of isolation for that test.
One more thing to try would see if he can borrow a factory tuned duplexer, with a similar frequency offset, and see if the results are any different.
I'm not sure what is meant by the GM300/Maxtrac radios experiencing desense, as they cannot tx/rx at the same time?
Hopefully you get to the bottom of the problem. Good luck,

Todd

Wavetar,

He uses a Q202 that can do better than 80 dB isolation at a 500 KHz split according to Sinclair’s info. He is using a 1 Meg split and is doing better than 80 dB as expected.
We run 100 watt Micor here with the same duplexer at the typical 2-meter band split of 600 KHz with no desense. There was desense at a time when there were antenna problems though.

Nand.

Nand wrote:I should point out here that John's problems shows up without any antennas at this point. He is using a 100 watt dummy load for all his tests and still has the problem. It wouldn't hurt to try a different dummy load though.

Nand.

So we stay on the same page, Nand indicated the problem
is there with a dummy load, and the person making the post
says there is No - Desense

kcbooboo wrote:

When the output of the duplexer is fed directly into a dummy load (NO coax), the receiver has NO desense problem.

I am assuming the poster is correct in saying there is No
Desense using a Dummy Load, and Nand may have just
overlooked his response.

Anyhow, this is only mentioned to insure we stay on the
same page due to the volume of nformation being offered.=======================================

Next Step:

Since the Desense changes with a increase / decrease
of the output power, your problem sure may be a strange
type of Antenna / Transmission line problem, or perhaps a
noise floor mix in the Power Amp, and or a Receiver enviroment
isolation problem.

I suspect the Rx Enviroment is the area that should be
given a close look. ( Especially the Shield covers, and DC
Power Input. )

One way to make this determination is to simply disable
your Source Transmitter !

Run a " External " Transmitter into your Filter / Duplexer
with perhaps a Maxtrac or some other radio with 40 Watts
of power, and make your measurements again.

Look at the Receiver while the External Transmitter is ON
feeding the Exsisting Filter Duplexer, and the feedline
connected.

Desense still there ? ( you now may be able to exclude the MSF transmitter ) You will have to make some jumpers for this
test, and RG-142B/U is needed.

RF energy floats around everywhere @ a radio site, and again,
your problem could be as simple as a Ground Cover not being
secure, or a bad Jumper Cable.

I have seen Bad Jumpers running from the Cabinet to the
Load source.

I am sure you are not using a cheap piece of cable !

Also, could be a Filter / Duplexer hardware failure. Just because
it works in a Dummy Load, does not give it a "clean bill of health"

A dirty transmitter, poorly grounded Filter / Duplexer, and even
the " RF BY-PASS capacitors along the DC source path can be the cause of your problem.

Since your radio runs on a Common DC Source, Examination
of the Power Supply distribution lines should be checked as well.

They provide a perfect path for picking up RF energy, and taking
it directly to the RX front end.

Since your operasting Tx and Rx Freqs are close, there is not
much room for error here.

A Excellent way of checking your receiver is to have a External
Transmit Source ( like with a small ground plane ) have your
Receiver tied into the main feedline with a Coupler so you can
run your Signal Gen coupled into the Rx Port.

With the Generator turned on feeding the Receiver through
a RF Coupler:

Turn on the External Transmitter, and bring the Antenna close ( DO NOT MAKE DIRECT CONTACT ) to all the Filters, Cavities, Feedlines, and what ever.....

See if you have a " External " Source to Receiver input issue.

Desense seen ?

Could very well point to a RF By-pass cap / filter issue.

If you notice Poor Receiver Performance with a External Transmit
Source ( Either internal or External ) you may have a problem with the Receiver in its operating environment.

Monty

Lets get this straight ones and for all.

Below is John’s input.

John Isella wrote: The test that I ran several weeks ago which really has me stumped is where the dummy load is on the output of the duplexer with whatever power output I want and the RX is as clean as can be, not any hint of de-sense, nothing at all.

The "problem" can be seen clearly on meter position 2 of the metering panel. The normal noise floor is one reading, when the transmitter is keyed with the dummy load at the output of the duplexer the meter does not move at all.

When a transmission line with dummy load or antenna is connected that meter reading increases indicating de-sense. If the squelch is open you can hear "typical" de-sense.

The above is the puzzling problem. An antenna is not involved at this point.

With the dummy load on the duplexer without a feed line in between, there is no desense at all.

Add the feed line between the duplexer output and the dummy load, and now there is desense.

Also from kcbooboo’s previous post,

kcbooboo wrote: All the points that have mentioned antenna systems are nice to know, but they have absolutely no relevance to the current problem, as John, Nand, and I have stated numerous times.

When the output of the duplexer is fed directly into a dummy load (NO coax), the receiver has NO desense problem.

When the output of the duplexer is fed through 100 ft of Andrews SuperFlex 1/2 inch Heliax into a dummy load, the receiver has 12-20 dB of desense.

Let me add, that if both the Heliax, connectors and the dummy load are without fault, then this scenario makes no sense.

Nand.

Hi:

I believe the collective efforts here are
directed to help in the diagnosis of the
problem reported, and therefore its truly
important ( as Nand has indicated ) the
information be clear.

if any offense was taken, it was not
intentional

============================

That being settled, I would like to offer
a few Tech Tips we have used in the field.

Inductive coupling with just a wire is not
a idea method ! Just the wire itself can
bring unwanted energy into the picture.

Couple of ways to make these measurements
properly.

Use a Feed through with a adjustable " RF " Tap.
to tie into active RF paths both transmit and receive.



If you do not have some of these RF Taps, I would recommend
taking a piece of RG214 about 3 ~ 6 inches long with good
type N connecotrs on each end, and cut a small access hole
( about 1/4 " ) in the Center. Carve out enough material to
have access to the Center Conductor if desired, and
then use a smaller piece of RG142B/U with a BNC on one end
and place the Teflon Center Conductor from the RG142B/U
( with the Center cut down as much as possible ) into the
Access hole of the 214. Solder the surrounding shileds
of both cables together.

This alows a complete feed through, along with a isolated
tap in the center to accomodate test Cables from a Service
Monitor or Analyzer.

DO NOT make direct Contact with the Center of the RG214 ! Just
bring it within 3/16-1/4 " of center. This should allow enough inductive coupling without having to much impact on the Cable being used.

If you have doubts about the distance, cut a small piece of the
RG142B/U Teflon Center, remove the center wire, and just cut
the Teflon to size. Place it into the access hole, then add the small
pig tail with the BNC on one end.

This gives you about 40~ 50 Db of attenuation, smaller the
distance to the Center of the RG214 , less attenuation.

Its used in cases where there is no other choices available. And
it does work.

When dealing with High Power Transmitters, its very important
to follow saftey proceedures

Anyhow, if needed, make 2 of them, use one for Rx and one for
TX and this makes it a lot easier to make measurements properly
and minimizes leakage.

Hopefully with all the assistance Nand and others have given,
you should be able to isolate the problem area, and correct the
problem

MS

Thanks again for all the input.

Another detail that I did not mention is that I have tried a Phelps Dodge bandpass filter on the RX, it's one of the ones that looks like a black mobile duplexer, nice sharp profile although it does not cover the TX frequency, it had no effect.

Jim202 thanks for the good ideas. This machine is here at my home and the antenna is just on the roof so no funny tower issues thankfully, then there is the situation where it has the de-sense with just a transmission line with the load on the other end and no antenna system at all. The duplexer is a Sinclair 4 cavity which looks in good shape, the T connectors were NG but they have been replaced with good ones. So I think the local RF environment is out.

Nand, good thoughts I was telling Bob earlier today that I have yet to try the SA TG across the RX TX ports with the different configurations on the output, very good idea which I have not tried, will do so likely later today. As you point out a bad circulator could do all sorts of things however keep in mind the different transmitters I have tried. I will try to find out if I have one in there. I have tried small different lengths of additional coax but will try some more.

Monty, good point on the need for a good SA, I have and am using a Tek 495P with matching tracking gen. Interesting on the torque requirements on the MSF, not surprising since they sort of tighten up slowly, I have them pretty tight by hand...

When I do drop the output power the de-sense does go down however never does go away, in fact there is still a good deal there with the power down to about nothing, as if there is a floor for the de-sense.

Bob, perhaps I can put the bandpass filter that I have between the RF deck and the power amp, this would have some chance of helping. Also I could try it on the output of the PA at low power. It's certainly not rated for 100w.

Yes there is actual de-sense on the air, when a mobile station is out driving around I can turn the TX on and off and realize a large improvement in RX signal due to lack of TX interference.


Using GM300 as a transmitter hooked into the TX port of the duplexer while the MSF receiver is still hooked to the RX side of the duplexer resulted in the same de-sense as using the MSF TX.

Thanks for the suggestion on using a different transmitter, I have already done that and get the same de-sense, in other words I have substituted the MSF transmitter with two different radios, a Radius and a GM300 and just keyed them up and see/hear the same de-sense on the MSF RX. This follows the same pattern where it is there when coax and dummy load is introduced in the output of the duplexer and gone when the dummy load is placed right on the antenna port of the duplexer.

Nand, I am taking steps to get another load to try.

Nand, the situation makes lots of sense, it just happens to be de-sense! )

Monty, thanks for the idea on making a tap I will put one together.

Just did another test, I placed the bandpass filter tuned to the TX freq in between the RF deck and the PA. No difference in de-sense. I then lowered the power output all the way down which was in the area of 5 watts and put the bandpass filter on the PA output. At this output level the de-sense is small but clearly seen on meter position 2 and can be heard on the RX and heard by radiating a small signal with a sig gen and an antenna. The de-sense was still there, I then brought the power up to 25 watts and the de-sense increased in line with what I had without the filter at 25 watt output. I took the filter out reduced power to compensate for small loss in the filter and measured de-sense was the same. So, this is perhaps a good test indicating that the spectral content of the TX is not a factor. Does anyone dis-agree? This was all done with antenna and transmission line by the way.

OK, so now we believe that there's no wide-band noise coming out of the MSF5000's transmitter. Unfortunately that bandpass filter may not have a steep-enough skirt to pass the TX frequency while rejecting the RX frequency, especially at a 1 MHz split.

I think the suggestion of a 5 MHz split is interesting, but that would put one side or the other out of the ham band. You could still use your signal generator as a source, and at least check the coax/no-coax situation at the output of the duplexer. Shouldn't take too much time to retune the duplexer. I'd move the TX freq so you don't have to retune the MSF5000's receiver.

John, I have one of those RF taps, as well as the 6-12 dB pads. Perhaps I should send them to you for further testing. If nothing else, you can use the pads and tell us if anything changes, tuning-wise, with the duplexer. If it makes no difference, and you'll have to keep good track of the tuning positions, then you can go on and continue to tune duplexers without them. But if it does make some difference, and the eventual outcome is better performance, then a slight thought-process adjustment will be required.

I have my RF tap set for 30 dB, so my modulation meter indicates 40 milliwatts of power coming from it when I have a MaxTrac transmitter putting out 40 watts into a dummy load. You could similarly adjust it using your signal generator and spectrum analyzer. The device is somewhat frequency dependent (i.e. it is NOT flat), so you'd need to calibrate it on your band (I did it at 446 MHz).

With all this stuff you should be able to settle some disputes and further eliminate pieces of the puzzle.

Bob M.

Monty wrote:Hi:

I believe the collective efforts here are
directed to help in the diagnosis of the
problem reported, and therefore its truly
important ( as Nand has indicated ) the
information be clear.

if any offense was taken, it was not
intentional

No offence taken at all Monty. I just wish that John had left any mention of the antenna out of the picture because it just confused the situation.

Both Bob and I have been conversing with John for quit sometime about this problem and in the end suggested that John gets on Batlabs because we have people here with many years of practical experience like you and others. Some of the suggestions made on the board were already investigated, but that is unavoidable. John tried a variety of radios and duplexers with the same similar results, I believe.

For me the most puzzling item is that the dummy load has a different effect on the tests if it is connected trough a length of Heliax instead of a direct connection. This would indicate that both the Heliax and the dummy load do not have the same impedance.

Nand.

HI:

In response, yes ! I do like to see the strange
ones which is why I responded.

I have seen some real brain twisters.....One
was " Rust " on the tower, and it had to be
Sand Blasted and Repainted to solve the
problems.

But that aside, I realize the Dummy Load vs Antenna
can be puzzling, and although the Cable has a variable
velocity factor the Dummy load is constant.

And these HIgh End Radios like a Constant Enviroment
They even like the same temp range as we do !

As you indicated use ofa alternate dummy load as a
A-B comparison was not a bad idea. Rare, but could
happen.

With all the input, I still feel pretty confident it's a
Receiver environment problem, since it was ruled
out earlier, and who know, where the RF is leaking
back into the RX Port.

A good RF Voltmeter properly isolated could tell '/ but
a good Spectrum Analyzer does the same thing.

Sometimes, just saturating the Front end of a receiver
input can be a issue.

Anyhow, glad you are helping them, and good luck
on the quest....Please post the findings once you have
located the problem.

MA

I read here that you have tried different TX units. By any chance have you tried a diifferent RX?

If you have the Radius and Maxtrac, try programming one of them up for a receiver. Cable it right to the duplexer. See if you get the same problem. You don't need to wire it in to key the repeater. Just use it to see if you get the desense. Manually key the MSF5000 transmitter for the test.

Right now it looks like changing one item in the system at a time, might be a way of closing in on the real problem. One small step at a time is how we all have solved our problems over time.

Jim

Hi Jim, Yes I have tried all this, in fact the off the wall aspect of this problem is that I have changed everything, that's everything, every element has been swaped out at least once in many cases 3 or more times, that's 3 duplexers, 4-5 different receivers, handheld, GM300, MSF. 3-4 transmitters, 4 different transmission lines, plain old RG-8, yes I know it's no good, LMR-400, 9913, Heliax. This is why I am here with this problem. It virtually has to be something else outside the hardware, my tuning techniques have workd for years so no reason to go astray now. I just don't know, have been working this problem for almost 6 months. The only thing I know is when a dummy load is on the output of the duplexer it's clean as can be, with transmission line in the middle I get de-sense. For the benifit of readers, prior setups were not reduced to this point, in other words I had not isolated it down to the load and the coax I was just dealing with de-sense on the air, only when I got the MSF-5000 did I really start to swap out EVERYTHING else. I did try the 3 different duplexers and other radios first however.

Thanks for reading about my mess...

John

HI:

With all the information now in.

Its my professional opinion you have
a RX - Filter / Receiver environment
problem !

I base this on the following:


Radio is Clean and works well with a Dummy Load.

The Dummy Load " shields " most of the RF that the
Transmitter generates, and thus, the MSF 5000 Filter
on the RX Port is not being saturated from the outside
since there is no antenna

A laymen explanation ( as best I can describe )

Have you ever gone to a Mountain top where there
is FM & TV Broadcast Stations? Ever wonder why your
Broad Band Handie-Talkie goes numb ?

Its due to over saturation of the front end !

I know it may sound dumb, but the torque settings
on the MSF5000 Covers are critical ! and as I mentioned
when RF is being generated into the air, what could be
happening is the RF is " Getting By " the Filter, and thus
ending up @ the RX receiver port.

( Your Spectrum Analyzer can varify this ) Use the
Narrow most settings and look what RF elements
are being seen at the Rx Port. Just disconnect the Rx
in the MSF5000, run it directly to the Analyzer.

Having a loose cover is like not having a Filter at all !
=======================================
Those of us who have worked on Mitreks learned a long
time ago, both transmit and receive performance improved
100% when the Pre-selector cover was removed, cleaned,
and reinstalled.

Many times, the improvement was enormous.=======================================

I do not have the torque settings at where I am at, but
I will list them.

If you have one of these Sturtevant/Richmont Cal 36-4
drivers, you can use that or a suitable substitute.

We use the above tool almost every day to insure radios
are in proper working order.



In short, RF energy is probably getting by the Duplexer / filter,
and no matter what you do, there is enough leakage to have a
impact on the receiver.

A " Bad Cable " can do exactly the same thing !

One main reason I only use RG-142B/U or RG-223/U when
it comes to Patch / Test Cables.

When using a single antenna for both transmit and receive,
its very important everything be perfect. Some over look that
issue. When using High Power, it becomes mandatory that everything be ideal.

Think of how techs deal with 250 and 375 Watt Transmitters ?

Anyhow, again, good luck in your quest.

Monty

At last some new different data from testing. This afternoon I set up to do the test with the SA and TG looking at the TX to RX port path of the duplexer under two conditions, one with the load at the output of the duplexer and one with the load at the distant end of the hard line. I set things up and took photos of both setups on the SA. Then prior to breaking down the setup with the hard line I did a transmit test to verify that there was de-sense in that configuration. There was no de-sense with the load at the end of the 100 feet of hard line. I determined this first by virtue of no increase in meter position 2 reading, to confirm this I powered up the sig gen and put the small antenna on the front panel to radiate. I can get enough signal in the air to get into the RX in spite of everything being terminated. With a nice noisy signal, and not moving my body which yes does change the "path loss" there was no de-sense. The best bad news I have had. I can not account for why this was the case or why my prior test had de-sense on the hard line test. I then put the repeater back on the air with the antenna on the roof, same old performance with 11-12 db de-sense. I then tried a second antenna which is located inside the attic with 9913 cable on it to see what would happen there. Slight power coming back, about 2-4 watts indicated out of 100 into the duplexer, nothing really considering it's in the attic. De-sense was much higher at this point, it took 22db increase in sig gen output to overcome the interference and the position 2 meter indication was similarly higher reflecting higher de-sense.

Now, this has me thinking along with Monty's last post that perhaps radiated RF is just getting into the RX directly outside of the duplexer/transmission line path. This is at least viable I think in light of the latest test with the hard line and load that had no de-sense. It would also account for the fact that in spite of the many changes I have made the de-sense is always there. Also with the second antenna in the attic it's all that much closer to the repeater and would induce a higher RF field at the repeater itself. I still can not account for the different test data with the load and hard line test. At least I now have some direction to go in.

Is your main antenna a real antenna [Celwave/Decibel/Antenna Specialists]???

Many people claim amateur antennas [Diamond/Comet] just won't cut it for duplex use.

You are starting to muddle things up John.

In order to get to the bottom of this, stick with the dummy load and find out why there is no desense when it is connected directly to the duplexer, and why it has desense when it is connected to the duplexer though the Heliax cable. Also now, why was there no desense in your last test with the load at the end of the cable. What did change or is eratic? Can you duplicate you findings.
Bringing an antenna in the picture is pointless at this time if you cannot be sure that it works correct under all conditions with a dummy load.

You also didn’t mention what the results were of your last tracking generator measurement form the TX port to the RX port. Was there a difference with the and without the Heliax in there?

Nand.

XMO, no it's not a real antenna.

It's a Diamond, I have tried 3 Diamond units, one mono-band, 2 dual band with the same results. And yes at some point I will consider buying a station master, that point might be tomorrow.

Nand, I am just as perplexed at this point as you imply. I can not explain the difference, I will do some more testing after this email. Yes I see your point on the antenna business, I can send you the photos, no difference in the two configurations, then again for what ever reason neither was there a difference in performance, no de-sense in either case, coax or no coax. So, no difference in the SA measurements may not mean a heck of a lot.

More to follow.

Ok, I just unrolled the Heliax and ran it around the house again to repeat the prior findings and there is no de-sense with the dummy load at either the duplexer output or at the end of the Heliax now. Needless to say this is a simple test so it is hard to imagine that I would have screwed it up when I originally got the Heliax just for this test. Having said that I have no idea why it would be different. Outside the fact that my prior test results made no sense these results provide a lot more to work with. As a matter of fact what prompted me to buy the Heliax was that I put the dummy load on the end of the coax on the roof with the actual transmission line disconnected from the antenna. At that point I thought I was onto the issue of using 9913/LMR400 shielding problem in duplex use since I had de-sense in that test. But, when I got the Heliax and tried it I still had the de-sense. Who knows, cockpit error can not be ruled out. At this point I think I need to run the Heliax up onto the roof and connect it up to the antenna and see what happens...

Can I post photos here? If so how?

JI

So it looks like your problem disappeared but can return anytime.

Just to be sure, inspect every connector. It is my experience and likely everyone else here that connector problems are 99% of the problem if not done by one self.

Pictures can be posted by surrounding the URL with [img] and [/img] without any spaces. The picture needs to be located on a server some place, perhaps your ISP’s server.

Click on the link below to see an example of this in Monty's post.

http://batboard.batlabs.com/posting.php ... e&p=106221

Nand.

Well, as far as the test results of de-sense with the Heliax and dummy load being an intermittent which exhibits the same characteristics as the installed coax and antenna, this is unlikely. I put the connectors on the Heliax myself so they should be good. I would be happier to accept the responsibility for conducting the test with de-sense with the Heliax and a load at the end in my dreams vs. another intermittent. I can not explain how this could be different... perhaps I have been banging my head against the wall a bit too much. What is not different is that I still have the same de-sense when hooked up to the antenna and on the air. If I hook the Heliax up to the antenna and it goes away this is certainly worth something. If this happens, perhaps tomorrow I can try this then I must be somehow at fault in the test of Heliax and dummy load with de-sense. If not then I have a worse mess with two problems that are very similar, this too is not unheard of as we all know.

I sent you the photos Nand, I don't have a way to host them but if you think it would somehow be usefull here feel free to post them.

Thanks,

JI

The photos can be seen here if my link stays up,

http://www.domtoren.com/Desense

or here if it doesn't.

http://web.ripnet.com/~wrom/Desense

By the way, I don't care much for your dummy load. I don't trust it.

Nand.

I see from the photos one item that I would get rid of as soon as possible. Unless my eyes are fooling me, it looks like you have some sort of a right angle type N connectors at the T antenna common point of the duplexer. I have never liked this type of connector. When I was working at the Mass. Institute of Tech. many years ago, we found these connectors were bad for maintaining 50 ohms. They also had a real bad SWR when they were used.

If it isn't too much of a problem, I would make some new cables and get rid of the right angle connectors. Use a regular straight type N connector. We also found that the center pin on these connectors were generally poorly soldered. The shield was hard to maintain a good solid connection. In general, they are a poor connector for repeater service in this case.

This may not make any change to your system, but it will eliminate at least 2 sources of potential problems.

Jim

Jim202, I agree, I am not fond of these connectors either, I have changed out all the cables with other random 100% shielded ones and it made no difference, but yes I should change them out.

Thanks for posting the photos Nand.

I do now recall something that is different in the test setup. I no longer have the known bad N Tee connectors on the duplexer cans. Out of 4 two of them were getting hot and one fell apart when moved around, these were original Sinclair parts with the labels on them, I have only two intact I think since I took two of them apart. The center conductors of these were not only a bad connection but they were not even an interference fit! Good ones have the center Tee conductor screwed into the other through pin center conductor, these had the Tee pin just banging around in a larger hole in the through pin. I will try to put the two remaining ones back in the TX side later and see what happens. Perhaps I have had two problems.

Nothing else changed for this test, I did not open up any of the rest of the repeater or make or break any other connections but the bare minimum, as far as I know...

I thought you changed all the duplexer's Tee connectors weeks ago, and were still having the desense problem?

Are you now telling us that with 100 ft of Heliax stretched out, and the dummy load attached to the far end which is now not sitting 5 feet from your MSF5000, that the desense problem has gone away? Or did I miss something in the translation? There has to be something different in the test setup where you had desense with the coax between the duplexer and dummy load vs the same setup minus the coax.

I would not be happy about having a 100 ft hunk of Heliax just sitting on the floor. At HF frequencies, they use coax loops for matching networks, so there must be something else going on in the viscinity of the coax that affects the signal carrying ability (impedance, mutual coupling, etc).

Maybe there's an alligator floating around outside who ate a 1 MHz rusty metallic connection acting as a mixer, and it is watching you through the window, but it only positions itself appropriately when it notices you've connected the coax to your duplexer !

Bob M.

Yes I did change the T's weeks ago, I also have not re-run the Heliax test since, or I don't think I have anyway, no test log here so I can't tell you exactly when I did what, who would think this would get so complicated to require an engineering review process...

The Heliax is not coiled up. I looked out side and don't see and alligators.

Did not do any testing today, need to redo some of the old tests not to mention hook the Heliax up to the antenna.

Apparently there was some mis-information, probably initiated by me, but uncorrected until I spoke with John on the phone yesterday. I apologize for the mis-guidance.

Sometime during the past 3 weeks, the desense problem experienced with the Heliax coax between the dummy load and duplexer disappeared. All the times that desense was being experienced was when the LMR400 with the outside antenna connected was attached to the duplexer. With that disconnected, everything is clean now and working just like you'd expect. Therefore, the LMR400 and/or antenna are causing some system imbalance that results in desense.

What I was told, and what I relayed to this group, was that he had desense with just a dummy load attached to the end of a new piece of Heliax. Well, he may have had it once, over 3 weeks ago, but since then a lot of things have been done to the equipment (new cables, new T connectors, eliminating "whiskers" in the receiver) and that setup is now clean as a whistle, with or without coax between the duplexer and the dummy load.

At one time, more than 3 weeks ago, a dummy load was attached to the end of the LMR400 and the desense still remained. All these past weeks when I was told there was desense with the Heliax in the line between the duplexer and dummy load was really 3 week old news; the test had not been done again until late this week and that configuration is now clean.

As an additional test, John put a 12dB pad between the receiver and duplexer and noted a lower meter-2 reading with the transmitter on, as well as decreased desense and overall sensitivity (naturally). With a 18dB pad, the desense was almost gone. With 30dB of attenuation, there was absolutely no desense at 100 watts transmitter power output. All of these tests were run with the LMR400 and outside antenna, currently the only condition thich generates desense now.

The next step would be to put a dummy load back on the end of the LMR400 to see if the coax is causing the problem or the antenna is. If the desense goes away when just the antenna is removed, that indicates one thing. If it remains, then the Heliax needs to be installed and the test repeated until there is no more desense with a radiating antenna connected to the system. The whole purpose of the Heliax was to replace the LMR400, so that would also be the logical next step if the coax is shown to be bad.

Earlier in this discussion people were questioning the antenna and coax, and Nand and I kept steering people away from that subject and back to the facts (as we knew them at the time, but now incorrect) that there was desense with coax between the duplexer and dummy load. Again, I apologize for insisting that the antenna and coax be ignored for the time being.

It would appear that everything in the shack (MSF5000, duplexer, cables, etc.) is now functioning as it should be. Now John has to look for a low level signal, near the input of the receiver, that's at a signal level of around -110 dBm; that should only take him a few hours to locate !

Bob M.

I don’t believe there is a need to apologize by you Bob, John clearly stated in his first post here (Nov 1) , that he had no desense with a dummy load and desense with a dummy load and feedline connected to the duplexer.
He seems to have been reporting desense for a feedline dummy load combination where he was using a LMR400 / antenna instead. Also, from previous emals he gave me the impression that he was testing with Heliax now and not LMR400 because it is pretty well condemned for repeater use.

John apparently had several problems at the same time. One problem was the defective T connectors on the duplexer and the other is the use of LMR400 and some antenna that is not suitable for repeater use either.

It was only on Nov 3 that he re-run the test with the Heliax and dummy load to find out that he fixed that problem several weeks ago, and now must realize that the other problem is with the antenna / LMR400 that should not have played a role in his test to start with, because he needed to prove the duplexer Heliax combination first.


Nand.